Optimization of geometrical parameters for two possible states with different material properties



I have a material with changing properties - 2 states. I want to optimize 3 geometric parameters so that the contrast ratio between these two states is maximal. Is it possible to somehow insert this into optimization? E.g. for the first state I need to optimize 4 parameters to get the highest transmission, and for the second state the lowest transmission.

Using a nested parameter sweep in an optimization
How can I access the results of a nested parametersweep in the custom figure of merit script (for optimization)?

Hi Milos,

It is possible to do something like this by using a sweep nested inside the optimization and creating your own figure of merit. In the nested sweep you run separate simulations for the two material states. The figure of merit must be a number calculated from the results of the two simulations. For instance, you could use the difference in transmission between the two states.

I have created a simple example of an AR layer to illustrate this: AR_twostate_optimization.fsp (238.7 KB)…

In the “AR structure” group you can set the thickness and material of the AR layer. I created two dielectric materials, “mat_state1” and “mat_state2”, as examples of two possible states for the layer. In the nested sweep you need to specify these states under parameters. For the results I chose the reflection measured above the structure.

In the optimization settings you need to create your own figure of merit, for example the difference between the two states: fom = R(1)-R(2).

Some useful KB links for optimization are the following:
Custom optimization algorithms
Nested sweeps

For more sophisticated optimization you can use MATLAB via Lumerical’s automation API.

How can I access the results of a nested parametersweep in the custom figure of merit script (for optimization)?

Thank you very much!


I’m also confusing with this problem. In sweep, there are two simulations, simulation for ‘mat_state1’ and ‘mat_state2’. Each simulation gives a reflection coefficient R. If I want to optimize the difference between these two R. How could I set the FOM. I find the fom=R(1)-R(2) is not used as Fiugre of merit in the setup—Fiugre of merit—yes part